Chile Earthquake
Felipe Toro, Felipe Rubilar, Matias A. Hube, Hernn Santa-Mara
and Tamara Cabrera
ABSTRACT
Approximately 300 bridges, representing less than 3% of the total number of Chilean
bridges, were damaged during the 2010 Maule earthquake. From these damaged bridges, which
include pedestrian bridges, 20 suffered collapsed spans. The objective of this study is to conduct a
statistical analysis of the 88 underpasses located in a representative section of the main route 5.
This analysis is aimed to correlate the bridge characteristics with the observed damage level. Past
studies reveal that this type of post-earthquake analysis is a robust tool to nurture the decisionmaking process for upgrading vulnerable structures and for guiding future performance based
engineering studies. To conduct the statistical analysis a database was built with the underpasses
characteristics such as location, presence or absence of reinforced concrete (RC) stoppers or
concrete lids, skew angle, and seismic zone. Additionally, the database includes the description of
damage, the required repairing action, and the estimated repairing cost. From this study it is
concluded that critical variables affecting the underpasses behaviour are the presence or absence of
stoppers, the skew angle, and the location of the bridge. Bridges with concrete lids should be
avoided because their repairing cost is higher than the repairing cost of bridges with RC stoppers
and have a larger probability of collapse The average repairing cost for the considered underpasses
is estimated in 11% of their construction cost and the skewed underpasses were more vulnerable
than non skewed bridges because their repairing cost is higher.
__________
Hernn Santa-Mara, Associated Professor, PUC, Vicua Mackenna 4860, Santiago, Chile
Matias A. Hube, Assistant Professor, PUC, Vicua Mackenna 4860, Santiago, Chile
Felipe Rubilar, Graduated Student, PUC, Vicua Mackenna 4860, Santiago
Felipe Toro, Graduated Student, PUC, Vicua Mackenna 4860, Santiago, Chile
Tamara Cabrera, Project Reviewer Engineer, MOP, Bandera 76, Santiago, Chile
1
INTRODUCTION
A Mw 8,8 Chile earthquake occurred on February 27th, 2010, in the central southern region
of Maule (Boroschek et al., 2012). More than a third of a million buildings were damaged to
varying degrees, including some cases of total collapse of major structures (Elnashai et al., 2012).
The transportation system registered 830 failures on roads in both the public and private
transportation networks. Of the nearly 12,000 highway bridges in Chile, approximately 300 were
damaged, including 20 with collapsed spans (Buckle et al., 2012). A total of 221 public bridges and
91 concession bridges suffered damage or collapse (MOP, 2010). The most common observed
damage modes were: the damage between super and substructure, transverse displacement and/or
rotation of the superstructure, unseating of spans due to skew, and cracking of prestressed concrete
(PC) girders induced by the impact with transverse stoppers. Detailed damage of Chilean Bridges
after the earthquake has been reported by several authors (Buckle et al., 2011; Elnashai et al., 2012;
Kawashima et al., 2011; Hube et al. 2010; Schanack et al., 2012; Yen et al., 2011). The total
economic losses due the earthquake has been estimated in 30 US$ billion (Elnashai et al., 2012),
which is equivalent to approximately 17% of the GDP of Chile. The cost to repair the damage in
the road infrastructure has been estimated at US$ 850 million (MOP, 2010). The emergency and
reconstruction program of the Ministerio de Obras Pblicas (MOP) considers about 500 US$
million which is intended to repair only the public road infrastructure.
The Chilean bridge structures are classified in four categories: underpasses, overpasses,
bridges, and pedestrian overcrossing. An underpass and an overpass are bridge structures where the
main route runs underneath or over the secondary route, respectively. The bridge category is a
structure that is used to cross rivers or creeks. The statistical analysis presented in this study is
conducted using the 88 underpasses of a section of Route 5 from the Maipo River, 29 km south of
Santiago, to the southern area of Chilln, 412 km to the south. Typical underpasses are two-spans
bridge structures with a central pier and two abutments, see Figure 1. In the considered section of
route 5 there are two concession companies involved in the construction and operation of the road;
the total number of bridge structures is 375, with 88 underpasses, 53 overpasses, 122 bridges, and
112 pedestrian overcrossing. The main objective of this study is to conduct a statistical analysis of
damaged and undamaged underpasses using a database of 88 structures to obtain the critical
variables that triggered the observed damage behavior. The second objective is to estimate the
repairing cost of damaged underpasses, which is of great relevance for future performance based
engineering studies.
After the arrival of the concession companies in the mid-90s, typical bridge designs for
underpasses have been evolving. Before the 90s bridge design consisted of simply supported PC
girders, with a reinforced concrete (RC) slab, transverse RC diaphragms at the supports, RC
stoppers, and vertical bars or hold downs for preventing uplift (Figure 2a). In the new design, the
RC diaphragm was eliminated, as shown in Figure 2b. Additionally, in some bridges the RC
stopper was replaced by a concrete cover, or lid, with a small gap between the PC girder and the
cover, as shown in Figure 2c. The structural difference between the latter two configurations is
explained later. It is important to note that in other damaged underpasses, not considered in this
study, the RC stoppers and the vertical hold downs were replaced by steel stoppers for preventing
both the transverse displacement and the vertical uplift (Hube and Rubilar, 2011). Most
underpasses considered in this statistical analysis have the structural configurations shown in
Figure 2b and 2c, without diaphragm.
the deck. The hold downs are designed with a vertical coefficient of Kv=Ao/2g neglecting the effect
of the dead load, and they must be at least 22 mm in diameter. However, the design code does not
specify a prestressing force for these hold downs. Transverse diaphragms are required to maintain
the geometric section of the deck. However, these diaphragms can be eliminated in seismic zones 1
and 2 if an adequate performance of the bridge can be demonstrated.
To limit transversal displacement of the superstructure, bridges are usually provided with
lateral stoppers. A stopper is a structural element located at the abutment or cap-beam and it is
located at the side of the exterior girder, or in some cases, in between the girders. A RC stopper at
the end of a cap beam for a bridge of the presented database is shown Figure 3a. According to the
design code (MOP, 2002), each stopper has to be designed for half of the total lateral seismic load
of the superstructure. Since the lateral seismic load of a simply supported superstructure is resisted
in most cases by two stoppers (one on each end of the superstructure), the pair of stoppers resist the
total seismic load. Additionally, the stoppers should be ductile enough to prevent unseating of the
spans, they should be taller than 30 cm, which is not the case of the stopper shown in Figure 3a,
and a minimum gap of 5 cm should be provided between the stoppers and the girders.
In the presented database, several underpasses were constructed without stoppers but with
concrete lids, as shown in Figure 3b. A concrete lid is a vertical element located at the edges of the
abutments or cap-beam and it is installed for esthetical reasons. This element has minimum
reinforcement and it is not designed to resist lateral force or deformation. When bridges are
designed with concrete lids, the transverse resistance is provided entirely by the elastomeric
bearings and the vertical hold downs. The hold downs prevent the uplift of the superstructure and
are intended to maintain the normal force in the bearing to prevent sliding. But, as mentioned
earlier, no prestressing force is specified for the hold downs, and sliding of elastomeric bearings
was observed in several bridges.
Shortly after the earthquake, MOP issued new design requirements that need to be revised
with further research. Some examples of the modifications incorporated in these requirements are:
transverse diaphragms are mandatory for all bridges, and each stopper has to be designed with the
total lateral seismic load, doubling the pre-earthquake design force.
a) RC stopper
b) Concrete lid
Figure 3. Difference Between RC Stopper and Concrete Lid
BRIDGE DATABASE
As the damage in overpasses was distributed through different highways along the country,
information of all underpasses located in a representative section of the main route 5 was
consolidated into one database. The database of 88 underpasses was constructed using structural
drawings, damage inspection reports and soil mechanics reports of most underpasses. This
information was obtained with the support of MOP. The database includes the following
information for each bridge: location, seismic zone, length and width of the superstructure, skew
angle, number of girders, presence or absence of RC diaphragm, presence or absence of RC
stoppers or concrete lids, number of vertical hold downs, description of damage, and the required
repairing action.
Based on the detailed description of damage, four damage levels (DL) were defined for the
whole bridge as well as for elements such as girders, stoppers or concrete lids, expansion joints,
vertical hold downs, and columns. The damage in the access roads and soil embankments in the
abutments was not considered in this study. The damage level for the whole bridge was defined
based on the operation condition immediately after the earthquake. DL-1 stands for a bridge with
no damage; DL-2 stands for a bridge that suffered low damage and required minimum or no repair;
DL-3 stands for a bridge with damage that required to be repaired; and DL-4 is assigned to a
collapsed bridge. A similar definition of the damage levels was used for the elements. Examples of
the four damage levels for the whole bridge and for the girders are shown in Figures 4 and 5,
respectively.
with an effective acceleration of 0.4g. The middle and east of the country are defined as seismic
zones 2 and 1, respectively, with effective accelerations of 0.3g and 0.2g, respectively. The
definition of these discrete seismic zones affects the design of bridges, where different discrete
values of the seismic coefficient are used at different seismic zones. The number of underpasses of
the database located in seismic zones 3 and 2 are 24 and 64 bridges, respectively. The percentage
of underpasses located in each seismic zone, with different bridge damage level is shown in Figure
9. This figure shows that the percentage of undamaged bridges (DL-1) is similar for bridges
located in both seismic zones. The three underpasses of the database that collapsed (DL-4) are all
located in seismic zone 2, which represents 5% of the bridges in this zone. It is important to note
that the collapsed bridges were designed with a lateral force 33% smaller than that of bridges
located in seismic zone 3.
Figure 10. Location of Damaged Bridges and Definition of Seismic Zones: Seismic Zone 3 shown
with Dark Grey and Seismic Zone 2 shown with Light gray
REPAIRING COSTS
The detailed repairing action for each underpass after the earthquake was obtained from
MOP. With this information we are able to correlate the damage level of each bridge with the
repairing cost.
The total cost to build a brand new underpass is estimated using unit price information
provided by MOP for a similar bridge. The construction cost of each underpass of the database was
estimated using the unit price information and the number of girders, the width of the
superstructure and the span length. The cost of the abutments and piers are also included, but the
cost of the soil embankments and the access roads are excluded from this analysis.
The repairing cost of each bridge was estimated using the required repairing action and the
damage level defined in this study. For the case of girder damage levels (see Figure 5), the
repairing costs associated to DL-1, DL-2, DL-3 and DL-4 were estimated as 0%, 5%, 50%, 150%
of the total cost of new girders, respectively. The cost of replacement of the expansion joints was
estimated as 300% of the cost of a new joint because it requires demolition and complex concrete
casting adjacent to the joint. The repairing costs of concrete lids and RC stoppers of abutments and
piers, and the repairing cost of superstructure alignment for bridges with displacement larger than
50 mm are also included. Finally, repairing costs associated to columns or foundation repairs were
not considered because the columns and foundations of the considered bridges were undamaged.
10
The estimated repairing cost of the underpasses considered in this study varies from 0% to
120% of the bridge cost, and the average repairing cost is 11%. A repairing cost of 120% of the
cost of a new bridge is assigned to bridges with damage DL-4 because of the demolition cost. The
average repairing costs for bridges with concrete lids and RC stoppers are 14% and 10%,
respectively. These results suggest that concrete lids should be avoided and that bridges should be
designed with stoppers. The average repairing cost for skewed (angle of skew larger than 5) and
non-skewed bridges are 26% and 7%, which means that skewed bridges are more vulnerable than
non-skewed bridges. Finally, the average repairing costs for bridges located in seismic zones 2 and
3 are 13% and 7%, respectively. The last result is consistent with the fact that bridges in seismic
zone 3 were designed with higher seismic demand.
CONCLUSSIONS
A statistical analysis of the 88 underpasses located is conducted in this research. The main
conclusions from this study are:
Bridges provided with RC stoppers did not collapse. However, 13% of bridges with RC
stoppers were undamaged, while 34% of bridges without RC stoppers (with concrete lids) were
undamaged.
Skewed bridges presented larger dispersion in bridge damage than non-skewed bridges. The
percentage of bridges with displacement or rotation of the superstructure was 26% and 13% for
skewed, and non-skewed bridges, respectively.
The discrete definition of the seismic hazard level in Chile influenced the bridge behavior. The
three bridges of the database that collapsed were located in a seismic zone with a lower
effective acceleration. There is an urgent need to develop a new and continuous seismic hazard
map in Chile.
The average repairing cost for the underpasses is estimated in 11% of their construction cost.
Bridges with concrete lids should be avoided because their repairing cost is higher than the
repairing cost of bridges with RC stoppers and have a larger probability of collapse.
The average repairing costs for skewed and non-skewed bridges are 26% and 7% of their
construction cost, respectively. Skewed bridges are more vulnerable than non-skewed bridges,
and that current Chilean design provisions should be revised to provide uniform vulnerability
to both type of bridges.
ACKNOWLEDGEMENTS
This research was supported by the Chilean National Commission for Scientific and
Technological Research (CONICYT) through projects FONDEF D10i1027 and FONDECYT
#11121581.
REFERENCES
American Association of State and Highway Transportation Officials, AASHTO (1996), Standard
Specifications for Highway Bridges, 16th edition, American Association of State and Highway
Transportation Officials, Washington, DC.
Boroschek, R.L., Contreras, V., Kwak, D.Y. and Stewart, J.P., (2012). Strong Ground Motion Attributes of
the 2010 Mw 8.8 Maule, Chile, Earthquake, Earthquake Spectra, Vol. 28, pp. 519-538.
11
Buckle, I., Hube, M., Chen, G., Yen, W., and Arias, J., (2011), Structural Performance of Bridges in the
Offshore Maule Earthquake of February 27, 2010, Earthquake Spectra, Vol. 28, pp. 533-522.
Elnashai, A.S., Gencturk, B., Kwon, O., Hashash, Y., Kim, S.J., Jeong, S. and Dukes, J., (2012), The Maule
(Chile) earthquake of February 27, 2010: Development of hazard. Site specific ground motions and
back-analysis of structures, Soil Dynamics and Earthquake Engineering Vol. 42, pp. 229-245.
Kawashima, K., Unjoh, S., Hoshikuma, J. and Kosa, K., (2011). Damage of Bridges due to the 2010
Maule, Chile, Earthquake, Journal of Earthquake Engineering, 15(7), pp. 1036-1068.
Hube, M., Santa-Mara, H. and Villalobos, F. (2010). Preliminary analysis of the seismic response of
bridges during the Chilean 27 February 2010 earthquake, Obras y Proyectos, Edicin 8, Primavera, pp.
48-57.
Hube, M.A. and Rubilar, F., (2012), Capacity Evaluation of Steel Stoppers of Reinforced Concrete Chilean
Bridges, The International Symposium for the 25th Anniversary of the CISMID, August 17-18, Lima,
Peru, Paper TS-2-2.
Instituto Nacional de Normalizacin, INN, (1996), Diseo Ssmico de Edificios, NCh433. Of. 1996,
Instituto Nacional de Normalizacin, Chile (in Spanish).
Ministerio de Obras Pblicas, MOP, (2002), Manual de Carreteras, Volumen N3, Instrucciones y Criterios
de Diseo, Captulo 3.1000, Direccin de Vialidad, Ministerio de Obras Pblicas, Chile (in Spanish).
Ministerio de Obras Pblicas, MOP, (2010), Cuenta Pblica 2010, cuenta sectorial, Ministerio de Obras
Pblicas,
Chile,
http://www.gobiernodechile.cl/cuenta-publica-2010/ministerio-de-obras-publicas/
cuenta-sectorial/ (in Spanish).
Schanack, F., Valdebenito, G. and Alvial, J., (2012), Seismic Damage to Bridges during the 27 February
2010 Magnitude 8.8 Chile Earthquake, Earthquake Spectra, Vol. 28, pp. 301-315.
Yen, W-H. P., Chen, G., Buckle, I., Allen, T., Alzamora, D., Ger, J., and Arias, J. G., (2011), PostEarthquake Reconnaissance Report on Transportation Infrastructure Impact of the February 27, 2010,
Offshore Maule Earthquake in Chile, Federal Highway Administration Publication No. FHWA-HRT11-030, Washington, D.C.
12